A Longitudinal Approach to Evaluating Therapist and Client Contributions to Alliance on Outcomes in Borderline Personality Disorder

Methods for studying therapeutic alliance have primarily examined the impact of the early alliance on outcomes. This does not allow for an understanding of the contributions of client, therapist. and dyadic factors to the alliance. Also, the alliance may change over time, highlighting the need for longitudinal methods. Efforts have been made to develop approaches that decompose the contributing factors and their impact on outcomes, but these findings may not apply to clients with borderline personality disorder (BPD). Our study extends previous research by replicating Baldwin et al.'s (2007) approach to disentangling therapist versus client-therapist dyad effects while using a time-varying framework. Participants (n = 156) were individuals diagnosed with BPD randomized to 1 year of dialectical behavior therapy or general psychiatric management. Outcomes were general psychiatric severity and interpersonal functioning measured at baseline and every 4 months. Client-rated alliance was measured at these time points. Early alliance predicted interpersonal functioning (p = .0132) with a significant contribution from clients (p = .0248) but not therapists (p = .2854). In the time-varying analysis, client contribution to the alliance was significant (p = .0022). For general psychiatric severity, client contributions to the alliance were significant (p < .0001) but not therapist contributions (p = .6779). Client contribution to the alliance was significant (p = .0168) in the time-varying model. Results suggest that in a BPD sample, time-varying alliance is a better predictor of rate of change in outcomes compared with the alliance measured at a single time point. In contrast to other studies, client, not therapist, contributions to alliance were significant in predicting outcomes.